> For the complete documentation index, see [llms.txt](https://serotolabs.gitbook.io/tradeclash/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://serotolabs.gitbook.io/tradeclash/game-design/bidding-strategy.md). # Bidding Strategy The auction has simple rules. The strategy is not simple. This page covers the mental models that separate a first-lot player from a hundredth-lot player. *** ## Reading Your Private Position Your hole card is the single most important piece of information in the auction. Before you bid on Street 1, ask: * **How much value does this position contribute?** If `(currentPrice - entry) × size` is high relative to the payout range, the lot is probably valuable. * **What category is it?** A strong position in a familiar category gives you more confidence. * **What's the size?** Large positions swing the trueValue more. A small position with big price movement matters less than a big position with modest movement. Your private position is a sample of one from a pool of seven. It's informative, not conclusive. *** ## Street 1 Bid Sizing You have the least information and the most uncertainty. Two approaches: **Conservative**: Bid low or pass ($0). Let others reveal their conviction. Use Street 1 bid data as a free signal for Street 2. **Aggressive**: If your private position is strong and you know the category, bid high. A 50% gap over the next-highest bid wins the lot immediately. The risk: you're bidding on one data point. Most experienced players lean conservative on Street 1 unless their private card is exceptional. The information you gain from seeing other bids is usually worth more than winning early. *** ## Reading Other Players' Bids After each street, all bids are revealed. This is the richest signal in the game. * **High Street 1 bid** → That player saw a strong private position. The lot is probably valuable. * **All low Street 1 bids** → Either everyone saw weak cards, or everyone is sandbagging. Context matters. * **A player who bid high on Street 1 then drops on Street 2** → The shared reveal changed their mind. The new shared position might be weak. * **Rising bids across streets** → The table is converging on higher value estimates. Be careful of herd behavior — the winner's curse loves crowds who agree. Bid patterns are behavioral data. Treat them like a poker tell, not a price signal. *** ## First-Place Wins Going for a first-place win is a calculated risk: **When it works**: You have a strong private position, you know the category well, and you're willing to risk overpaying for the reward of early victory. The 50% gap on Street 1 is steep — you need real conviction. **When it backfires**: You gap the table by 50%, win the lot, and the reveal shows the trueValue was $180 on your $310 bid. Winner's curse at its worst. **Street 2 first-place wins** (20% gap) are more defensible. By then you have your private card, one shared card, and everyone's Street 1 bids. The picture is clearer. Rule of thumb: First-place wins should be rare and deliberate. If you're going for them every lot, you're donating to the leaderboard. *** ## The Winner's Curse You will overpay. It's not a question of if — it's a question of how often and by how much. **Managing it**: 1. **Anchor to the payout range, not your private position.** Your card might be the best one in a weak lot. 2. **Discount your estimate.** If you think the lot is worth $300, bid $250. The gap protects your downside. 3. **Pass on weak lots.** $0 is a valid bid. The best players pass more often than they win. 4. **Track your hit rate.** If you're winning 40%+ of lots, you're probably overbidding. Healthy win rates are closer to 20-30%. The leaderboard rewards cumulative P\&L, not win count. A player who wins 2 lots at $40 profit each outranks a player who wins 5 lots and nets -$10. *** ## Information Cascading Each street adds information and changes the math: | Street | You know | Decision quality | | ------ | ------------------------------------------ | ---------------------------------- | | 1 | Private position + payout range | Low — one data point, wide range | | 2 | + 1 shared position + all Street 1 bids | Medium — triangulating | | 3 | + 1 more shared position + all bid history | High — but burn cards still hidden | The cascade means Street 3 bids are the most informed. But they're also the most crowded — everyone has the same shared information. Differentiation comes from: * Your unique private position (only you have it) * Your read of other players' bid behavior * Your category expertise * Your risk tolerance *** ## Phase 2 Preview: Creator vs Bidder When player-created lots launch, the game gains a second dimension. Creators bundle positions and set payout ranges — their edge is lot construction. Bidders evaluate and price lots — their edge is valuation. Creator profit = winningBid - trueValue (when bidders overpay) Creator loss = trueValue - winningBid (when bidders get a deal) Building good lots and bidding well on them are different skills. The best players will do both. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://serotolabs.gitbook.io/tradeclash/game-design/bidding-strategy.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.